mohammadali-khazen / body-orientation-detection-model

The body orientation detection model identifies the field-of-view of a worker on construction sites.

The body orientation detection module uses signal blockage by a human body to identify an approximate worker's body orientation on job sites. In fact, a receiving beacon mounted on the worker's chest cannot capture signals from the reference transmitting beacons located behind the worker. The angle of the detected transmitting beacons by the chest-mounted receiving beacon to the worker's estimated location can provide the worker's field of view. To this end, in-lab experiments were carried out to collect data between the chest-mounted receiving beacon and reference transmitting beacons for eight ordinal orientations. Then, a DNN model was trained to predict the body orientation, using signal data and geometrical relationships between the beacons.